kolinko

23 ms·

571.
by kolinko 3y ago
Interesting thought. I think ground-based GPS wouldn't be too difficult though - we already have most of earth covered by GSM/3G/LTE, and with updated towers you could have something as precise, if not more, as GPS. Of course
572.
by kolinko 3y ago
doesn’t recommend ≠ recommends against
573.
by kolinko 3y ago
doesn’t recommend ≠ recommends against
574.
by kolinko 3y ago
Many users, myself included, choose iOS because it’s free of viruses and malware - that is the benefit of the review process. If you don’t want that, you can use Android.
575.
by kolinko 3y ago
Consumers (myself included) already made their decision by choosing an iOS. I don’t want every other game dev/streaming service pushing me to install their own “store” like they do on the other platforms. That’s why I keep choosing App
576.
by kolinko 3y ago
And not for you? Where do you live that you would possibly be positively affected by such change?
577.
by kolinko 3y ago
As someone from central-easter Europe - yes, please and thank you!
578.
by kolinko 3y ago
There is another thing - low res games were played on a CRT monitor that had a natural smoothing of pixels. If you launch any old game on any modern monitor it will look way worse than it used to be.
579.
by kolinko 3y ago
I wanted to say that they are to smart to expect dmca to protect them. But then, I think that surely they would use copyright to block competition from using their model directly.
580.
by kolinko 3y ago
T&C yes, but not copyright. This is fully consistent with them opposing copyright and not opposing paywalls/api limitations.
581.
by kolinko 3y ago
T&C yes, but not copyright.
582.
by kolinko 3y ago
Most people have that suspicion, myself included. But the API models stay verifiably same, and also OpenAI would have little reason to lie about not updating ChatGPT. Otoh, I can easily imagine helper/censorship models and chat’s syste
583.
by kolinko 3y ago
Well, FLOP make sense, kind of - you can make computing as cheap as you want, but without algorithmic improvements the FLOP count will stay the same. And algorithmic improvements will probably be discrete - happening once in a while, not co
584.
by kolinko 3y ago
The fear is that you won’t get better model from training on synthetic data from a worse model, and you will miss a lot of modern day knowledge. Imho the first is not necessarily so, but the second will be a real hindrance. You can talk abo
585.
by kolinko 3y ago
Depends on what you want to use AI for - for a smarter AI we could probably do with the open source material, but better training techniques and some innovative breakthroughs in the model. Also, an interesting side note is that certain coun
586.
by kolinko 3y ago
It sounds similar to the fears media rang about tat you can find a bomb making tutorial on the Internet - people were genuinely afraid of that. (Otoh an agi can bring unforseen consequences for humanity - and that’s a genuine fear)
587.
by kolinko 3y ago
No, it's practice. The model doesn't change day by day. Kind of like saying "let's watch the same youtube video every day and see if it changes" - it won't, that's not how youtube works, there is no point
588.
by kolinko 3y ago
Sure, but with a system like this you still need to build for reusability of a rocket - if not then even a smaller rocket will be more expensive than what SpaceX is planning to do with Starship. And if you're building a fully reusable
589.
by kolinko 3y ago
Oh, nice, TIL. Good point about surface to volume ratio.
590.
by kolinko 3y ago
I asked GPT to continue the conversation :) quantumconfusion 6 hours ago | root | parent | next [–] True, but there's also the flip side. People might inadvertently adopt AI-like patterns in their writing, especially as AI becomes mo
591.
by kolinko 3y ago
It doesn't work like that, not with text. If you make a text model smart enough it will be able to generate whatever you want.
592.
by kolinko 3y ago
> these large companies also want to keep copyright protection intact, for their regular business activities Care to share an example? I didn't hear of OpenAI or anyone else arguing or trying to sue anyone for abusing the copyright.
593.
by kolinko 3y ago
For music there are already similar mechanisms in place in many countries - in Poland it's ZAiKS, in US it's ASCAP. They collect fees from organisations playing copyrighted music publicly. (I agree that it would be terrible if the
594.
by kolinko 3y ago
Hacker News was always about people sharing their project ideas, that’s why there is a whole Show HN section here.
595.
by kolinko 3y ago
He speaks very unclearly, instead of saying GPT-4-turbo he says 4.5 preview. 4.5 is invention of his. Also mixtral medium - no idea of what he means by that. Not to mention a claim that mixtral is as good as gpt-4. It’s on the quality of gp
596.
by kolinko 3y ago
Yes! And there was a second mini series connected to this one! Yakamoz S-245
597.
by kolinko 3y ago
Launching from altitude gives you very little, so does spinlaunching. It is because it’s about speed, not altitude. Witch Spinlaunch, they plan to laumch around 5000kmph, and you need a speed of around 25000kmph to reach orbit. In other wor
598.
by kolinko 3y ago
They will probably have an energy storage facitilty nearby. There’s little benefit to bundling these, and the cost of maintenance is increased. Also, batteries will have shorter timespan than the turbines. It may begin to make sense when ba
599.
by kolinko 3y ago
But then, the same can be said of a computer with a finite-sized memory. I think the real thing in the paper is the computational complexity of calculations. With turing machines, each step is O(1), with transformers it's at least O(n)
600.
by kolinko 3y ago
It was the same in communist countries, even more so. If anything, we’re having less of this now that we switched to capitalism than before.

More ›